Evaluating an IoT under-mattress sensor mat for detecting anomalies in sleep parameters: A pilot study

Sleep is equally important as a healthy diet, and exercising and not getting adequate sleep can lead to being less physically active during the day. Still, sleep is underrated by many people. Measuring sleep at home using conventional methods is not practical, given the complexity of equipment, so several user-friendly sleep monitoring devices are continually being developed. This paper aims to evaluate the effectiveness of an IoT under-mattress sensor mat for identifying anomalies in sleep parameters under a real-life scenario out of the lab. The sensor mat embodies a microbend fiber optic sensor that is sensitive enough to capture the mechanical activities caused by the heart. Five older adults participated in the study, and raw sensor data were gathered over several weeks. Limits of agreement and boxplots were employed to identify anomalies in three sleep parameters, i.e., wake-up time, bedtime, and time in bed. Also, a sleep history metric was applied to captures an aggregate measure of sleep behavior. Albeit using a single-channel sensor, sleep anomalies were detected and verified by the participants' caregivers.